How One Agency Uses AI Without Falling Into the Hype Trap

Every agency is suddenly an “AI-powered” agency. Most are just slapping ChatGPT into their workflows and calling it innovation. Taylor Thomson’s approach is more deliberate.

At WITHIN, Thomson leads cross-functional projects with data science teams to develop internal databases using generative AI technologies like GPT-4 and Bard. But the work focuses on specific operational problems rather than chasing headlines about artificial intelligence. His measured approach to technology implementation contrasts sharply with the hype cycle dominating agency marketing.

“We’re able to invest in best in class technology to help us be more efficient,” Thomson explains. The emphasis is on efficiency gains from solving actual problems, not implementing AI for its own sake.

One application involves client satisfaction surveys. WITHIN achieves over 50% quarterly response rates—unusually high for B2B contexts. The company uses AI to analyze open-ended responses at scale, identifying patterns that would be difficult to spot manually. This creates actionable insights rather than just satisfaction scores. Taylor Thomson’s role at the agency’s Denver headquarters involves overseeing these technology integrations while ensuring they deliver measurable value.

Another area involves competitive intelligence. Using Pathmatics (now part of Sensor Tower), WITHIN tracks social media spend across channels. AI helps process this data to identify trends—which competitors are increasing investment in specific channels, what types of creative are gaining traction, where market dynamics are shifting.

These aren’t flashy applications. They’re operational improvements that compound over time.

Thomson also recognizes AI’s limitations. “Technology alone doesn’t create results,” he notes. WITHIN’s tech stack includes Salesforce, Outreach, and various other tools. But Thomson emphasizes that value comes from systematic usage, not just tool acquisition.

The same principle applies to AI. Having access to GPT-4 doesn’t magically improve operations. You need clear use cases, disciplined implementation, and measurement of actual outcomes.

This measured approach contrasts with agencies making grand claims about AI transformation while changing little about their actual processes. Thomson focuses on specific problems where AI provides genuine advantage: processing unstructured feedback at scale, identifying patterns in competitive data, automating routine analytical tasks. His perspective on technology and finance emphasizes practical implementation over theoretical possibilities.

For agencies considering AI implementation, Thomson’s example suggests starting small with concrete use cases rather than announcing sweeping AI strategies. Find operational bottlenecks where AI actually helps. Measure whether it’s working. Expand gradually based on results.

The companies that get AI right won’t be the ones with the boldest announcements. They’ll be the ones that quietly solve operational problems while competitors are still figuring out what “AI-powered” means. Taylor Thomson’s professional background in both finance and operations positions him uniquely to evaluate which technology investments deliver returns versus which just generate press releases.